-
Is there a way I can use your QMIX implementation to train a model on a different environment? For simplicity assume the generic `env` variable has the same functions/attributes, etc as SMAC. I'm more…
-
The error is like this:
'Process Process-3:
Traceback (most recent call last):
File "/home/a/anaconda3/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File…
-
Hi,
this project is really awesome and the codes is well structured!
_**Is your feature request related to a problem? Please describe.**_
I have run some codes in examples/tutorials, but can n…
-
I've alread posted this question on stackoverflow but since i didn't got an answer there i will repost it here (https://stackoverflow.com/questions/61523164/ray-rllib-qmix-doesnt-learn-anything)
I …
-
When running `python3 src/main.py --config=qmix_smac --env-config=sc2 with env_args.map_name=2s3z`, it has a error in Ubuntu18.04:
> pysc2.lib.remote_controller.ConnectError: Failed to connect to the…
-
### Describe your feature request
The QMix example and the default model only uses a simple MultiDiscrete observation space. It would be nice if more abstract observation spaces and custom mode…
-
### What is the problem?
When trying to run QMix on a multi-agent environment, the QMixTorchPolicy object throws an error upon initialization, saying
```
Failure # 1 (occurred at 2020-06-01_14-46-…
-
### What is the problem?
Hi, I want to try QMIX, with 2 agents, each agent receives an observation `gym.spaces.Box(0,1, (3,))`, and the env_state (the one fed to the mixing network) is the concaten…
-
Hello everyone,
I wanted to implement a random policy masking the non-allowed actions, modifying this implementation (https://chuacheowhuan.github.io/RLlib_rand_policy/).
My code for `compute_ac…
-
I run the example code of qmix and it seems that it cannot learn good policy. The reward is nearly random policy.